Golf video browsing system using three stage audio segmentation

Content-based audio classification and segmentation is a basis for further audio/video analysis. In this paper, we present our work on audio segmentation and classification which employs support vector machines (SVMs). Five audio classes are considered in this paper: silence, music, background sound, pure speech, and non- pure speech which includes speech over music and speech over noise. A sound stream is segmented by classifying each sub-segment into one of these five classes. We have evaluated the performance of SVM on different audio type-pairs classification with testing unit of different- length and compared the performance of SVM, K-Nearest Neighbor (KNN), and Gaussian Mixture Model (GMM). We also evaluated the effectiveness of some new proposed features. Experiments on a database composed of about 4- hour audio data show that the proposed classifier is very efficient on audio classification and segmentation. It also shows the accuracy of the SVM-based method is much better than the method based on KNN and GMM.

In this paper, we describe an automatic- speaker based- audio segmentation and identification system for broadcasted news indexation purposes. We specifically focus on speaker identification and audio scene detection. Speaker identification (SI) is based on the state of the art Gaussian mixture models, whereas scene change detection process uses the classical Bayesian Information Criteria (BIC) and the recently proposed DISTBIC algorithm. In this work, the effectiveness of Mel Frequency Cepstral coefficients MFCC, Linear Predictive Cepstral Coefficients LPCC, and Log Area Ratio LAR coefficients are compared for the purpose of text-independent speaker identification and speaker based audio segmentation. Both the Fisher Discrimination Ratio-feature analysis and performance evaluation in terms of correct identification rate on the TIMIT database showed that the LPCC outperforms the other features especially for low order coefficients. Our experiments on audio segmentation module showed that the DISTBIC segmentation technique is more accurate than the BIC procedure especially in the presence of short segments.

Identifying different audio segments in videos is the first step for many important tasks such as event detection and speech transcription. Approaches using Mel-Frequency Cepstral coefficients (MFCCs) with Gaussian mixture models (GMMs) and hidden Markov models (HMMs) perform reasonably well in stationary conditions but do not scale to a broad range of environmental conditions. This paper focuses on the audio segmentation in broadcast soccer videos into audio classes such as Silence, Speech Only, Speech Over Crowd, Crowd Only and Excited, with an alternative feature set which is simplistic as well as robust to changes in the environment conditions. Support Vector Machines (SVMs), Neural Networks and Random Forest are used for the classification. The accuracy achieved with SVMs, Neural Networks and Random Forest are 83.80%, 86.07%, and 88.35% respectively. The proposed features and Random Forest classifier are found to achieve better accuracy compared to the other classifiers.

For TV and radio shows containing narrowband speech, Speech-to-text (STT) accuracy on the narrowband audio can be improved by using an acoustic model trained on acoustically matched data. To selectively apply it, one must first be able to accurately detect which audio segments are narrowband. The present paper explores two different bandwidth classification approaches: a traditional Gaussian mixture model (GMM) approach and a spline-based classifier that categorizes audio segments based on their power spectra. We focus on shows found in the DARPA GALE Mandarin training and test sets, where the ratio of wideband to narrowband shows is very large. In this setting, the spline-based classifier reduces the number of misclassified wideband segments by up to 95% relative to the GMM-based classifier for the same number of misclassified narrowband segments.

Focus on Authors

In rarefied gas flows, discontinuity phenomena such as velocity slip and temperature jump commonly appear in the gas layer adjacent to a solid boundary. Due to the physical complexity of the interactions at the gas-solid interface, particularly in the case of systems with local nonequilibrium state, boundary models with limited number of parameters cannot completely describe the reflection of gas molecules at the boundary. In this work, the Gaussian mixture (GM) model, which is an unsupervised machine learning technique, is employed to construct a statistical gas-solid surface scattering model based on the collisional data obtained from molecular dynamics (MD) simulations. The GM model is applied to study Couette flow for different inert gases (Ar and He) confined between two parallel infinite gold walls at different temperatures. A direct comparison between the results obtained from the GM model and the Cercignani-Lampis-Lord (CLL) scattering kernel against the MD collisional data in terms of the distribution of the predicted postcollisional velocities, and accommodation coefficients has shown that the results from the GM model are an excellent match with the MD results outperforming the CLL scattering kernel. As an example, for He gas, while the predicted energy accommodation coefficient by the CLL model is more than two times higher than the MD predictions, the value computed by the GM model is in excellent agreement with the MD results. This superior performance of the GM model confirms its high potential to derive a generalized boundary condition in systems encountered with highly nonequilibrium and complex gas flow conditions.

Summary We evaluated the effect of home telehealth on weight maintenance after a group-based weight loss programme. The home telehealth intervention comprised telephone counselling and home Internet-enabled digital video recorders (DVRs) with three channels of video programmes. The video content provided reinforcement and support to promote problem solving, prevent relapse and sustain motivation. Eighty-eight obese or overweight African-American women were randomized to receive monthly telephone counselling (control) or the home telehealth intervention. The weight change during maintenance was not significant in either group (0.6 kg in the intervention group, 0.0 kg in the control group), and there was no significant difference between them. Changes in diet, physical activity, social support and self-efficacy during the maintenance period did not differ significantly between groups. DVR use was low: during the intervention, the number of valid DVR viewings ranged from zero to 42 per person. DVR use was positively associated with previous attendance at the weight loss classes. Home video-based telehealth is a new method of delivering a weight loss maintenance intervention to African-American women. It had no effect on weight maintenance in the present study.

Validation of a digital audio recording method for the objective assessment of cough in the horse

To understand how people really interact with digital TV and digital video recorder (DVR) systems, 10 people were recruited, 6 with DVRs and 4 with non-DVR digital cable or satellite TV systems. Video-recording equipment was set up in their homes to record their TV viewing. The person viewing as well as the TV content being viewed were recorded as the person watched TV. Afterwards, the recordings were reviewed and analyzed by the experimenter together with the viewer, who then described his or her behavior. The results show that overall the most common way people interacted with TV was in avoiding advertisements and finding new programming to watch when the current show ended. These results can be used to find ways to improve how users experience TV systems.

This paper reports results from a video ethnographic study of TV advertising exposure for 22 individuals in eight UK homes with Personal Video Recorders (PVRs). PVR usage varied widely both within and between homes. For 21 of the 22 individuals, the PVR was used ? if at all ? as a backup when there was nothing on viewers' favourite channels which they wanted to watch live. As a result, of the 3480 individual opportunities to see commercials during the study, 70% were viewed live and only 30% were time-shifted. In combination with other, complementary, studies these results suggest that the impact of PVRs on advertising exposure will be limited. The paper also explores a number of other audience attitudes and behaviours.

Due to economic and ecological reasons the energy consumption of consumer electronics products like TVs and set-top boxes (STBs) is of increasing importance. As nowadays also Personal Video Recorders (PVRs) are widely used, the question is raised, how time-shifted TV affects the energy use of such devices. In this paper a framework for analyzing the impact of recording and later playback of TV content on the energy consumption of a PVR is presented. The underlying method allows taking the viewing habits of users into account. The framework is used to analyze the energy consumption of a recommender-based PVR that, based on personalized recommendations, automatically records TV content for the user. Further, a study of an approach for an energy efficient PVR design, which allows deactivating peripheral PVR components when not needed, is analyzed w/r/t potential energy savings.

Chapter 5 - Storage in Home Consumer Electronics Devices

Digital video recorders (DVRs) for closed-circuit television (CCTV) commonly have an in-built capability to export stored video files to optical storage media. In the event that a DVR is damaged, its contents cannot be easily exported. This renders the forensically-sound recovery of proprietary-formatted video files with their timestamps from a DVR hard disk an expensive and challenging exercise. This paper presents and validates a technique that enables digital forensic practitioners to carve video files with timestamps without referring to the DVR hard disk filesystem.

A PC-based stereoscopic imaging and measurement system (SIMS) was developed for undersea applications. The SIMS is capable of real-time image capture, recording and stereoscopic display as well as on-fine/on-demand 3-D measurement. The SIMS consists of a pair of underwater cameras, a controller, a pair of digital video recorders (DVRs), a PC to house the CPU, DVRs, hard drives, and other peripherals, and an autostereoscopic display. The cameras can be either mounted on a remotely operated vehicle (ROV) or carried by a diver to capture underwater scenes. Stereoscopic images are viewed on the glasses-free autostereoscopic display. Software modules were developed and/or adopted from published computer codes to operate the SIMS. Operations include those described above, camera calibration, and on-line/on-demand 3D measurement of features on the target objects. A virtual 3D camera calibration scheme was extended from a camera calibration method which captures the images of a coplanar checkerboard at several perspectives. The performance of the SIMS prototype for 3D measurement was evaluated by comparing the measured and actual distances between the checker corners on the checkerboard and between markers painted on a fish silhouette at several angles of orientation.

By applying the characteristic of stroboscopic effect, a new method named Stationary Counting Method for the calibration of the frame rate of digital video recorder has been developed at the Hong Kong Standards and Calibration Laboratory(SCL). The proposed method can determine the frame rate of digital video recorders with the use of a special synchronous counter designed by the SCL. The frequency signal input to the counter is traceable to the Cesium Beam Frequency Standard kept in SCL, as well as traceable to the SI unit. The target devices include high speed digital cameras, smartphones, closed circuit televisions (CCTV) and car cameras. It is common for vehicles to install car cameras to capture the moment when unexpected incidents occur. The recorded videos can be used as court evidences for accident investigation. Therefore, the accuracy of the frame rate is extremely important. The proposed method can be used to find the time duration of the video recorded by a digital camera, as well as the occurrence of skipped or extra frames of the captured video.